A month after the fresh-faced Facebook Oversight Board published 17 recommendations for how the platform could improve its content moderation, Facebook has responded with a resounding “OK, but it depends.”
One of the more significant changes Facebook will adopt relates to , to clarify that “health-related nudity” is allowed. The Oversight Board, which is an independent entity that rules on knotty content cases across all of Facebook’s platforms, recommended that Facebook update Instagram’s community guidelines around adult nudity to clarify that some nudity — if it is health-related, photos of breastfeeding, giving birth, breast cancer awareness, gender confirmation surgery, or in an act of protest — is allowed on the platform. Facebook agreed. It will take a while for that to go into effect holistically, but Facebook says it’ll provide progress updates. It’s not necessarily a win for the #FreeTheNipple movement, but it’s at least a step towards nuance.
Some nudity — if it is health-related, photos of breastfeeding, giving birth, breast cancer awareness, gender confirmation surgery, or in an act of protest — is allowed.
Including the Instagram nudity policy change, Facebook is acting on 11 of the board’s recommendations, “assessing feasibility” on five of them, and is dismissing one of them. This comes after the board on its first cases, which were immediately implemented.
Some of the other areas where Facebook says it is “committed to action” appear to be largely commitments to transparency. Facebook said it would clarify its community standards to include how it treats COVID-19 misinformation that could cause immediate physical harm and how it handles humor, satire, and personal experiences. Facebook is also launching a transparency center to help users better understand the platform’s community standards.
A recommendation Facebook won’t be implementing is one in which the board intriguingly asked for less oversight regarding COVID-19 misinformation. The board recommended that Facebook should “adopt a range of less intrusive measures” when users post information about COVID-19 treatments that contradict the advice of health authorities.
“In consultation with global health authorities, we continue to believe our approach of removing COVID-19 misinformation that might lead to imminent harm is the correct one during a global pandemic,” Facebook said in a statement.
A few of the board’s recommendations had to do with automation tools that make content moderation decisions. Facebook said it’d work to ensure its algorithms don’t automatically remove nudity posts that it does allow by refining its systems and sampling more training data. The board also recommended that users should be able to appeal decisions made by automated systems and ask that they are re-reviewed by an employee. That’s still under consideration as Facebook assesses its feasibility. Finally, the board recommended that users are informed when automation is used to make decisions about their content, and Facebook said it would test “the impact of telling people more about how an enforcement action decision was made.” Facebook is still considering how doable it is to disclose how many photos were removed automatically and how many of those decisions were then reversed after humans checked the algorithms’ work.
Facebook is also tackling some details of its Dangerous Individuals and Organizations Community Standard, the policy that captures everything from human trafficking to terrorist rhetoric. In response to a U.S. user posting a quote from Joseph Goebbels, the Reich’s minister of propaganda in Nazi Germany, Facebook removed the post for violating this policy.
The Oversight Board recommended that Facebook explain to the user which community standard it was enforcing when a post is removed and give examples of why it goes against that standard. Facebook agreed, adding that it will increase transparency around the standards by adding definitions of “praise,” “support” and “representation” over the next few months, since the Dangerous Individuals and Organizations Community Standard removes content that expresses “support or praise for groups, leaders, or individuals involved in these activities.” Facebook is also “assessing the feasibility” of providing a public list of “dangerous” organizations and individuals that are classified under the Dangerous Individuals and Organizations Community Standard.
These responses give us interesting insight into how Facebook will interact with the Oversight Board, which is still getting its sea legs after its first report. Only the board’s individual content decisions are binding, which is why Facebook had some wiggle room in its responses to these broader recommendations.
The 20-member board, which includes a Nobel Prize winner, academics, digital rights advocates, and a former prime minister, is also on if Donald Trump’s ban is permanent.